The Role of Illness Representations and Metacognitive Beliefs on Mood in People with Type 2 Diabetes: A study from the COVID-19 Period

ABSTRACT

Objective:

Diabetes poses a significant global health threat due to its increasing prevalence and associated comorbidities, straining healthcare systems worldwide. Depression and anxiety are frequent comorbidities in individuals with diabetes, and their underlying mechanisms have been a subject of research. This study investigated the influence of metacognitions and metacognitions about symptom control (MaSC) on depression and anxiety in people with type 2 diabetes (T2D), specifically examining their effects beyond illness representations.

Method:

Conducted during the COVID-19 pandemic, this study involved 153 T2D patients. Participants completed the Hospital Anxiety and Depression Scale (HADS), Brief Illness Perception Questionnaire (B-IPQ), Metacognitions Questionnaire-30 (MCQ-30), Metacognitions about Symptom Control Scale-Revised (MaSCS-R), and the Protection Motivation Theory Scale for COVID-19 Preventive Behaviors.

Results:

Hierarchical regression analysis showed that after controlling for socio-demographic variables, fear of COVID-19, and illness representations, negative metacognitions about symptom control predicted 2% of anxiety and 4% of depression; metacognitions predicted 10% of anxiety and 9% of depression.

Conclusion:

Metacognitions predicted both depression and anxiety, supporting the Self-Regulatory Executive Function (S-REF) model. Metacognitive therapy could be an alternative for managing emotional distress in T2D patients and may contribute to future research.

INTRODUCTION

Diabetes is defined as a chronic, broad-spectrum metabolic disorder requiring continuous medical care in which the organism is unable to use carbohydrates, fats, and proteins adequately due to insulin deficiency or impaired insulin action (Bastaki 2005, Adeghate et al. 2006, Turkish Society of Endocrinology and Metabolism 2020). According to international data, one in 11 people (nearly 463 million) is diagnosed with diabetes, and this number is expected to reach 538 million in 2030 and 700 million in 2045 (International Diabetes Federation-IDF 2019). Type 2 diabetes (T2D), which accounts for 90-95% of all people with diabetes, is defined as a disease characterized by insulin resistance, usually occurring in overweight or obese individuals, with onset after the age of 30 (World Health Organization-WHO 2019). The prevalence of T2D has been linked to the transition to a modern lifestyle, and its prevalence in children and adolescents has been reported to be increasing (Bloomgarden 1996, National Diabetes Consensus Team 2019). Data from Turkey showed a 15% prevalence of T2D in the Mediterranean region (Satman et al. 2013). It is predicted that the prevalence of diabetes among individuals aged 25 and above will reach 31.5% by 2025 (Sözmen et al. 2015). These prevalence rates confirm that diabetes is a serious public health problem in Turkey and internationally, and places a significant burden on health systems with comorbidities (Coşansu 2015).

Type 2 diabetes (T2D) is frequently associated with mental health problems such as depression and anxiety, and research shows that the prevalence of depression in individuals with T2D is approximately twice that observed in the general population (Ali et al. 2006, Roy & Lloyd 2012). The prevalence of clinically depressive symptoms in these patients is reported to be around 30% (Ali et al. 2006, Barnard et al. 2006), and the prevalence of major depressive disorder is reported to be around 10% (Egede 2004, Moussavi et al. 2007). Furthermore, as both physical symptoms and concerns about the progression of the illness increase, diabetes can be associated with anxiety (Goldbacher & Matthews 2007). A systematic review reported that 14% of people with diabetes had generalized anxiety disorder (Grigsby et al. 2002). Another study demonstrated that the lifetime prevalence of anxiety in people with diabetes was 19.5%. This rate was 20% higher than in people without diabetes (Li et al. 2008). In a parallel study, 48% of patients with T2D experienced depression, and 70% experienced anxiety (Rehman & Kazmi Farhana 2015). However, this rate was found to be 40.3% for depression and 31.4% for anxiety in outpatients with T2D (Ganasegeran et al. 2014). Studies show that this emotional distress associated with diabetes is predicted by helplessness (Akın 2013), being female (Alzahrani et al. 2019), being single, low socioeconomic status, and long duration of diabetes (Özbek 2014, Alzahrani et al. 2019). On the other hand, it has also been reported that a stressful environment can exacerbate emotional distress in people with diabetes (Mukhtar & Mukhtar 2020). The coronavirus disease 2019 (COVID-19), which originated in Wuhan, China, when the current study was planned, and spread globally, has also caused stress, leading to increased uncertainty and anxiety in patients with T2D (Alessi et al. 2020). While one study showed that 31% of T2D patients were concerned about COVID-19 infection (Sacre et al. 2021), another study found that the fear of COVID-19 was highest in people with diabetes with moderate depression (Basit et al. 2021). Since these findings suggest that the mental health of T2D patients may be affected by COVID-19, COVID-19 fear, and susceptibility scores were considered in determining the predictors of depression and anxiety levels in the current study.

The emotional distress experienced by individuals living with chronic diseases can be explained by individuals’ negative appraisals of illness, as posited by the cognitive paradigm (Cook et al. 2015a). Through this approach, illness representations have become a significant focal point within the Self-Regulatory Model (Leventhal et al. 1992, Leventhal et al. 2005). Illness representations, which include cognitive and emotional appraisals of illness type, duration, consequences, causes, and perceived controllability (Diefenbach & Leventhal 1996), are considered important variables predicting patient mood. Most studies suggest that perceiving low levels of illness symptoms and expecting fewer illness outcomes are associated with positive health outcomes, such as lower anxiety and depression, whereas perceiving high levels of illness symptoms and perceiving illness as chronic are associated with low vitality and functioning (Hagger & Orbell 2003). However, these studies have been criticized for their inability to explain the mechanisms underlying the relationship between illness representations and health outcomes (Hagger & Orbell 2003). Some researchers have proposed that the absence of a causal influence of illness representations on health outcomes is understandable. They have posited that not all individuals with chronic physical illness experience distress despite negative illness representations; emotional distress may arise when individuals respond to such thoughts with intense anxiety (Cook et al. 2015a). In this regard, the Self-Regulatory Executive Function (S-REF) Model, which serves as the foundation for metacognitive therapy, presents an alternative perspective (Cook et al. 2015a, Purewal & Fisher 2017, Heffer-Rahn & Fisher 2018). According to the model, psychological problems arise with cognitive attentional syndrome (CAS), which is a repetitive way of thinking that prevents the distortion or regulation of experiences from the metacognitive model. CAS is a three-way process consisting of perseverative thinking (e.g., worry, rumination, doubting), attentional strategies (e.g., focusing on symptoms, monitoring for negative thoughts and feelings), and unhelpful coping strategies (e.g., reducing activity, suppressing unwanted thoughts and images) (Wells & Mathews 1994). According to the model, CAS is activated and maintained by metacognitive beliefs. These beliefs can be classified into two distinct categories. While “positive metacognitive beliefs” are defined as those that correspond to the perceived benefits or advantages of engaging in cognitive activities that create and sustain CAS, “negative metacognitive beliefs” pertain to perceptions regarding the uncontrollability, significance, importance, and potential danger of thoughts and cognitive experiences (Wells & Mathews 1994, Wells 2009). Although this model has been extensively employed in elucidating the underlying mechanisms of psychological disorders, it has recently begun to be investigated in the context of physical diseases such as Parkinson’s (Allott et al. 2005) and chronic fatigue syndrome (CFS) (Maher-Edwards et al. 2012), and metacognitive beliefs and emotional distress were found to be related in these samples. Research in this area has shown that, after controlling for illness representations, metacognition contributes to explaining additional variance in psychological symptoms in cancer (Cook et al. 2015a), diabetes (Purewal & Fisher 2017), and MS (Heffer-Rahn & Fisher 2018) patients. In diabetes, metacognitive beliefs are reported to affect depression and anxiety by increasing rumination (Cherry et al. 2023).

Metacognition studies have explicitly focused on metacognitions for symptom control (MaSC) in chronic physical illness. A study utilizing the Metacognitive Profiling Interview (Wells, 2000) found that individuals with CFS focus on their symptoms, which negatively impacts their mood, hinders goal achievement, and exacerbates symptoms (Maher-Edwards et al., 2012). However, MaSC has not yet been tested in people with T2D. Therefore, this study aims to assess the predictors of depression and anxiety in T2D patients with current approaches, considering MaSC in the COVID-19 period. In this context, the study is expected to enrich the literature on metacognition, provide an alternative perspective on diabetes management, and enable interventions to be varied. In this direction, the following hypotheses (H) were tested.

H1. Metacognitions and MaSC will explain additional significant variance in anxiety levels of T2D patients after controlling for socio-demographic variables (gender, history of psychiatric diagnosis, monthly income, age), measures of COVID-19, and illness representations.

H2. Metacognitions and MaSC will explain additional significant variance in depression levels of T2D patients after controlling for socio-demographic variables (gender, history of psychiatric diagnosis, monthly income, age), measures of COVID-19, and illness representations.

METHOD

Participants

The participants were reached through Endocrinology Outpatient Clinic of Bursa Uludağ University Hospital (71.2% of the sample), and the “Type 2 Diabetes” Facebook closed group (28.8% of the sample). The following inclusion criteria were considered in the study: i) being between 18 and 65, ii) having a T2D diagnosis by a medical doctor, iii) having T2D for at least six months, and iv) volunteering to participate in the study. Individuals with a diagnosed psychiatric disorder were excluded from participation in the study.

A total of 162 participants were reached for the study, but people with psychiatric comorbidity (n=8) and people thought to have given random answers (n=1) were excluded. Finally, the data from a total of 153 participants were analyzed.

Study Instruments

Sociodemographic and Medical Information Form

This form consists of questions prepared by the researchers to obtain information on the socio-demographic characteristics (such as age, gender, income status) and medical conditions of the participants (diagnosis time, type of treatment). The form also includes questions about participants’ experiences with COVID-19.

Hospital Anxiety and Depression Scale (HADS)

The scale was developed to identify risk groups by screening for anxiety and depression in patients with physical illness (Zigmond & Snaith 1983). A 4-point Likert-type scale consists of a total of 14 items, seven of which assess depression and seven of which assess anxiety symptoms. A score of 8-10 on each subscale is characterized as mild, 11-14 as moderate, and 15-21 as severe depression/anxiety (Stern 2014). The item correlations for the anxiety subscale ranged between 0.41 and 0.76, while the item correlations for depression ranged between 0.30 and 0.60. A validity and reliability study of the Turkish version of the scale was carried out by Aydemir and colleagues (1997) on both patients and university students. The reliability values were found to be α=0.74 for anxiety and α=0.69 for depression in university students. The reliability values were found to beα = 0.74 for anxiety and α = 0.70 for depression in patients with physical illness (Aydemir et al. 1997).

Brief Illness Perception Questionnaire (B-IPQ)

The scale is a shortened version of the Illness Perception Questionnaire (IPQ) (Weinman et al. 1996) developed to assess the perceptions of patients with physical illness (Broadbent et al. 2006). The scale comprises nine items, eight assessing individuals’ explanations about the illness, and one assessing causal factors. The scale is scored between 0 and 10, except for the ninth item, which evaluates causal factors. The score obtained from the scale corresponds to how threatening the illness is perceived by the individual or how comfortable the individual is with the illness. Test-retest reliabilities for each scale statement ranged from 0.42 to 0.75 (Broadbent et al. 2006). The Turkish form consists of eight items, and a Turkish translation of the scale was conducted with cancer patients (Karataş et al. 2017). As in the original, the scale consists of two subscales, cognitive and emotional illness representations. The Cronbach alpha values are 0.85 for the total scale, 0.83 for the emotional illness representations, and 0.80 for the cognitive illness representations (Karataş et al. 2017).

Metacognitions Questionnaire-30 (MCQ-30)

The MCQ-30 was developed to assess various metacognitive beliefs and processes thought to play a role in the mechanisms of developing and maintaining psychological disorders (Wells & Cartwright-Hatton 2004). The 4-point Likert-type scale, ranging from ‘1 = strongly disagree’ to ‘4 = strongly agree’, consists of 30 items. It consists of five subscales: positive beliefs about worry, negative beliefs about uncontrollability of thoughts and danger, cognitive confidence, beliefs about the need to control thoughts, and cognitive self-consciousness. The total score obtained from the MCQ-30 varies between 30 and 120. The scale has good reliability and validity, with Cronbach’s coefficient alphas for the subscales ranging from 0.72 to 0.93. The Turkish version of the scale was carried out by Yılmaz and colleagues (2008) on university students and adults. Consistent with the original scale, a five-factor structure was obtained. The Cronbach alpha coefficient of the scale is 0.87 for the total scale, 0.89 for positive beliefs, 0,80 for negative beliefs, 0.89 for lack of cognitive confidence, 0.73 for need to control thoughts, and 0.80 for cognitive self-consciousness (Yılmaz et al. 2008).

Metacognitions about Symptom Control Scale-Revised Form (MaSCS-R)

It was the shortened version of the Metacognitions about Symptom Control Scale (MasCS) (Fernie et al. 2015), which was developed to assess beliefs about symptom focus, rumination, and worry in people with CFS (Fernie et al. 2019). The eight-item MaSCS-R was created by removing the items that did not fit the theoretical rationale or were considered difficult to understand (Fernie et al. 2019). The psychometric properties of this scale were assessed in individuals diagnosed with CFS (English), fibromyalgia (German), and type 1 diabetes (T1D, Lebanese). MaSCS-R consists of two subscales: positive metacognitions about somatic hypervigilance (P-MASH) and negative metacognitions about the uncontrollability and physical repercussions of engaging in symptom focus, rumination, and/or worry (N-MUR). The internal consistency coefficients for each factor of the scale’s English, German, and Arabic versions are acceptable and good (α=0.70-0.80). However, the internal consistency coefficient (α = 0.60) of the N-MUR subscale in the Arabic version was considered controversial. The Turkish translation of the MaSCS-R was carried out in people with different chronic physical illnesses and two similar sub-dimensions were obtained (Acar & Sertel Berk 2024). The internal consistency coefficients of the scales were 0.83 for the P-MASH and 0.66 for the N-MUR. (Acar & Sertel Berk 2024).

Protection Motivation Theory Scale for COVID-19 Preventive Behaviour

This scale, developed to assess the protection motivation against COVID-19, measures seven sub-dimensions: perceived severity, perceived susceptibility, fear, self-efficacy, response cost, response efficacy, and protection motivation (Acar & Kıcalı 2022). The fear and perceived susceptibility sub-dimensions of this scale were used in this study. Both fear (1=not at all, 5=very much) and perceived susceptibility (1=strongly disagree, 5=strongly agree) sub-dimensions are rated on a 5-point Likert scale (Acar & Kıcalı 2022).

Procedure

The data for the study were collected via face-to-face interviews with individuals who had applied to the hospital and via Google Forms from a Facebook group entitled “Type 2 Diabetes Group” between 2020 and 2021. Ethical approval of the study was obtained from the Clinical Research Ethics Committee of the Faculty of Medicine at Bursa Uludag University (date: 18.03.2020, decision number: 2020-5/28). Following the completion of medical controls, patients with T2D were informed about the details of the study, and their eligibility for participation and their diagnoses were verified based on data from the patient record file. Those who participated in the study online were informed that they could access online support if required. Each participant completed the demographic form and the following scales in a different order.

Data Analysis

All analyses were run on SPSS 20. Frequency, outlier (according to skewness, kurtosis values, and histogram), and missing data analyses were performed to test normality. Independent T-tests were done to explore the differences in the main variables according to socio-demographic variables and illness characteristics. The Pearson correlation test was used to assess the relationship between variables. Multiple hierarchical regression analysis was used to identify predictors of anxiety and depression. Hierarchical regression analyses tested the prediction that metacognitions would explain additional variance in T2D patients when controlling for age, gender, presence of a previous psychiatric illness (history of psychiatric disorder), average monthly income, fear of COVID-19, and illness representations. Before the main analyses, categorical variables (gender, history of psychiatric diagnosis, and the average monthly income) were transformed into continuous variables (dummy variables). Also, face-to-face and online data were compared for study variables. It was observed that the scale scores of the online and face-to-face data sets differed only for the lack of cognitive confidence (t(151) =-2.11, p=0.04). However, this finding was disregarded because of the low level of statistical significance, and the data sets were analyzed together. When the relationship between the predictor variables and the predicted variables was analyzed using the linear curve test in the fit line dot plot, it was found that there was a linear relationship. Only the predictors that showed significant zero-order correlations with the outcome variables were entered into the regression equation. Also, the multicollinearity was tested, and the tolerance and VIF values were within the appropriate range. These values are given in Table 3. Enter method was used in hierarchical regression analyses, and predictor variables were entered in the following order: Step 1; age, gender, history of psychiatric diagnosis, monthly income, and fear of COVID-19, Step 2; illness representations (B-IPQ total score), Step 3; metacognitions about symptom control (N-MUR), Step 4; metacognitions (positive beliefs, negative beliefs, cognitive confidence, need to control thoughts, cognitive self-consciousness subscales of MCQ-30).

Table 3.

Final hierarchical regression model for anxiety and depression in patients with Type 2 diabetes

	Anxiety								Depression
	Model 1 Summary						Multicollinearity		Model 1 Summary						Multicollinearity
	Adj. R2	ΔR2	B	Sth. Error	β	t	Tolerans	VIF	Adj. R2	ΔR2	B	Sth. Error	β	t	Tolerans	VIF
Demographic variables	0.21***	0.24							0.17***	0.20
Gender			0.54	0.48	0.07	1.14	0.84	1.19			-0.05	0.56	-0.01	-0.09	0.84	1.19
History of psychiatric diagnosis			1.76	0.61	0.18	2.88**	0.89	1.12			1.03	0.71	0.09	1.45	0.89	1.12
Average monthly income			0.88	0.97	0.06	0.91	0.90	1.11			3.06	1.13	0.17	2.70**	0.90	1.11
COVID-19 fear			0.03	0.07	0.03	0.39	0.63	1.58			0.01	0.09	0.01	0.11	0.63	1.58
Age			-0.06	0.02	-0.18	-2.80**	0.86	1.16			-0.01	0.03	-0.02	-0.32	0.86	1.16
Illness represantations	0.40***	0.18							0.37***	0.20
B-IPQ total			0.12	0.03	0.31	3.97***	0.56	1.80			0.16	0.04	0.34	4.32***	0.56	1.80
Metacognitions about symptom control	0.41*	0.02							0.41**	0.04
N-MUR			-0.11	0.12	-0.07	-0.86	0.45	2.23			0.14	0.15	0.08	0.96	0.45	2.23
Metacognitions	0.50***	0.10							0.49***	0.09
MCQ-30 positive beliefs			0.02	0.06	0.02	0.27	0.61	1.64			0.11	0.07	0.11	1.48	0.61	1.64
MCQ-30 negative beliefs			0.33	0.07	0.45	4.70***	0.37	2.70			0.25	0.08	0.29	3.06**	0.37	2.70
MCQ-30 lack of cognitive confidence			0.03	0.06	0.04	0.55	0.63	1.59			0.15	0.06	0.17	2.28*	0.63	1.59
MCQ-30 need to control thoughts			-0.09	0.07	-0.10	-1.24	0.54	1.86			-0.13	0.09	-0.12	-1.54	0.54	1.86
MCQ-30 cognitive self-consciousness			0.01	0.07	0.01	0.20	0.57	1.76			-0.12	0.08	-0.11	-1.48	0.57	1.76

^*p< 0,05,

^**p< 0,01,

^***p< 0,001

Adj. R2: Adjusted R-squared, ΔR2: Change in R-squared, VIF: Variance Inflation Factor, Std. Error: Standard error, COVID-19: Coronavirus Disease 2019, B-IPQ: Brief Illness Perception Questionnaire, MCQ-30: Metacognitions Questionnaire-30, N-MUR: Negative metacognitions about the uncontrollability and physical repercussions of engaging in symptom focus, rumination, and/or worry (cognitive attentional processes).

RESULTS

Descriptive Findings Related to the Sample

The sample comprised 86 (56.2%) women and 67 (43.8%) men. Participants were mainly in the middle and upper middle age group (M=50,66, SD=10,19). Sample characteristics are summarized in Table 1.

Table 1.

Demographic characteristics of the sample

Variables	N	%
Marital status
Single	10	6.5
Married	126	82.4
Divorced	6	3.9
Widower	11	7.2
Living together
With Spouse/Partner	31	20.3
With Spouse and children	93	60.7
With children	14	9.2
Alone	6	3.9
Other	9	5.9
Presence of children
Yes	134	87.6
No	19	12.4
Education
Illiterate	4	2.6
Literate	6	3.9
Primary school graduate	44	28.8
Secondary school graduate	14	9.2
High school graduate	32	20.9
University graduate	48	31.3
Postgraduation	5	3.3
Work status
Working	61	39.9
Not working	92	60.1
Reason for not working
Retired	42	45.6
Due to my illness	9	9.8
Other	19	20.6
Average monthly income*
0-1603 TL	9	5.9
1 604-5 200 TL	88	57.5
5 201-10 000 TL	46	30.1
10 001 TL ve above	10	6.5
Family income level
Low	34	22.2
Low middle	12	7.8
Middle	98	64.1
High middle	5	3.3
High	4	2.6

^*Income ranges according to 2020 data. TL: Turkish liras

The duration of T2D diagnosis of the participants ranged between 1-26 years (M=9,94; SD=6,54). 59.5% of participants were using antidiabetic drugs, 23.5% were using insulin, 14.4% had previously used antidiabetic drugs but were now using insulin, and 2% were following a diet alone. A total of 70.6% of the participants had another chronic illness, while 29.4% did not have another chronic illness. Among participants with comorbidities, 22.9% have blood pressure, 6.5% have thyroid disease, and 7.8% have cardiovascular disease. Also, 79.7% of participants did not have a COVID-19 test, 16.3% of participants had a COVID-19 test, which was negative, and 4% had a positive COVID-19 test. A total of 95.4% of participants did not have a COVID-19 diagnosis, and 87.6% did not experience the loss of a loved one due to COVID-19.

Results of the Comparisons of the Variables According to Sociodemographic and Illness Characteristics

A comparative analysis of depression and anxiety scores according to socio-demographic and illness variables indicated significant differences for only three variables: gender (t (151) = 2.13, p<0.05), history of psychiatric disorder (t (151) = 2.56, p<0.05), and average monthly income (F (3, 149) = 5.58, p<0.05). Women with T2D (M= 7.62, SD = 3.92) had more anxiety than men with T2D (M=6.34, SD = 3.33). The depression level of T2D patients with a history of psychiatric disorder (M = 7.74, SD = 5.30) was higher than that of T2D patients without a history of psychiatric disorder (M=5.42, SD=4.02). Also, the anxiety level of T2D patients with a history of psychiatric disorder (M=9.15, SD = 5.17) was higher than T2D patients without a history of psychiatric disorder (M=6.61, SD=3.17). Bonferroni comparisons showed that the depression levels of the T2D patients whose income was between 0-1603 Turkish liras (TL) (M=10.89, SD=2.89) were significantly higher than those of the patients whose income was between 1604-5200 TL (M=5.87, SD=4.04) and 5201-10000 TL (M=4.69, SD=4.55).

The relationships between the study variables (age, fear, susceptibility to COVID-19, depression, anxiety, illness representations, metacognitions, and MaSC) are given in Table 2. Depression and anxiety had low and moderate correlations with many other variables.

Table 2.

Relationships between the study variables

	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17
1.Age	1	0.36**	-0.24**	-0.09	0.08	0.07	0.05	0.07	0.09	0.02	-0.26**	-0.17*	-0.20	0.00	-0.01	-0.30	-0.04
2. Diagnosis time		1	-0.07	-0.12	0.08	0.05	0.08	0.01	0.10	0.09	-0.10	-0.10	-0.05	0.01	0.08	-0.27**	-0.15
3. HADS anxiety			1	0.70**	0.45**	0.23**	0.58**	0.30**	0.24**	0.28**	0.59**	0.20*	0.56**	0.15	0.39**	0.05	0.33**
4. HADS depression				1	0.51**	0.32**	0.57**	0.45**	0.28**	0.21*	0.59**	0.30**	0.50**	0.04	0.47**	0.01	0.32**
5. MCQ-30 Total					1	0.75**	0.82**	0.67**	0.76**	0.69**	0.45**	0.05	0.48**	0.37**	0.66**	-0.05	0.37**
6. MCQ-30 positive beliefs						1	0.44**	0.38**	0.49**	0.47**	0.22**	0.00	0.25**	0.26**	0.32**	-0.08	0.16*
7. MCQ-30 negative beliefs							1	0.49**	0.53**	0.52**	0.52**	0.13	0.51**	0.25**	0.66**	-0.06	0.35**
8. MCQ-30 lack of cognitive confidence								1	0.36**	0.14	0.32**	0.09	0.31**	0.20*	0.47**	-0.05	0.28**
9. MCQ-30 need to control thoughts									1	0.51**	0.33**	0.02	0.37**	0.34**	0.51**	0.06	0.31**
10. MCQ-30 cognitive self-consciousness										1	0.23**	-0.10	0.33**	0.35**	0.44**	-0.06	0.25**
11. B-IPQ Total											1	0.48**	0.86**	0.07	0.45**	0.05	0.45**
12. B-IPQ cognitive												1	-0.02	-0.05	0.18*	0.07	0.10
13. KHAÖ emotional													1	0.10	0.41**	0.02	0.46**
14. P-MASH														1	0.36**	0.10	0.21**
15. N-MUR															1	0.04	0.46**
16. COVID-19 susceptibility																1	0.17*
17. COVID-19 fear																	1

^*p< 0,05,

^**p< 0,01

COVID-19: Coronavirus Disease 2019, HADS: Hospital Anxiety and Depression Scale, MCQ-30: Metacognitions Questionnaire-30, B-IPQ: Brief Illness Perception Questionnaire, P-MASH: Positive metacognitions about somatic hypervigilance, N-MUR: Negative metacognitions about the uncontrollability and physical repercussions of engaging in symptom focus, rumination, and/or worry (cognitive attentional processes).

Predictors of Anxiety and Depression in T2D Patients

The hierarchical regression results for both H1 and H2 are summarized in Table 3. The results showed that the rate of prediction of anxiety by the control variables was 24% (F (5, 143) = 9.03, p<0.001). In this step, having a previous psychiatric disorder (β=0.18, p<0.05), fear of COVID-19 (β=0.28, p<0.001), and age (β= -0.23, p<0.01) predicted anxiety significantly. The addition of illness representations to the model resulted in an 18% increase in anxiety (F (1, 142) = 44.57, p<0.001). History of psychiatric disorder (β=0.15, p<0.05) and illness representations (β=0.50, p<0.001) predicted anxiety in this step. The addition of MaSC to the model resulted in a 2% increase (F (1, 141)=4.01, p<0.05). Gender (β=0.13, p<0.05), history of psychiatric disorder (β 0.14, p<0.05), illness representations (β=0.46, p<0.001), and N-MUR (β=0.15, p<0.05) predicted anxiety in this step. The addition of metacognitions to the model resulted in a 10% increase (F (5,136) = 5.80, p<0.001). The findings indicated that the overall model predicted 50% of the total variance. In the final model, anxiety was found to be significantly predicted by four variables: history of psychiatric disorder, age, illness representations, and MCQ-30 negative beliefs.

The results of the analyses of depression showed that the rate of predicting depression by control variables was 20% (F (5, 143) = 6.98, p<0.01). In this step, monthly income (β=0.23, p<0.05) and fear of COVID-19 (β=0.30, p<0.001) predicted depression significantly. When illness representations were added to the model, the increase in predicting depression was 20% (F(1, 142) = 47.78, p<0.001). In this step, monthly income (β=0.19, p<0.01) and illness representations (β=0.53, p<0.001) predicted depression significantly. When MaSC was added to the model, this increase was %4% (F (1, 141)=10.03, p<0.01). Monthly income (β=0.18, p<0.01), illness representations (β=0.46, p<0.001), N-MUR (β=0.24, p<0.01) predicted depression in this step. The addition of metacognitions increased 10% of the total variance (F (5, 136)=5.45, p<0.001). The findings demonstrated that the overall model predicted 49% of the total variance. In the final model, depression was found to be significantly predicted by four variables: average monthly income, illness representations, negative beliefs, and lack of cognitive confidence subscales of the MCQ-30.

DISCUSSION

The present study examines the predictors of depression and anxiety in patients with T2D, controlling for age, gender, history of psychiatric disorder, monthly income, and fear of COVID-19. This is the first study to consider MaSC in addition to metacognitions in T2D patients in the COVID-19 period. The results showed that the model of the predictors of anxiety in T2D patients was significant. Thus, MaSC predicted 2% of anxiety, and metacognitions predicted 10% after controlling for illness representations, sociodemographic variables, and fear of COVID-19. The final model indicated that 50% of the total variance in anxiety was predicted by four variables: history of psychiatric disorder, age, illness representations (total score of B-IPQ), and the negative beliefs subscale of MCQ-30. Similarly, results for depression demonstrated that the model was significant and MaSC explained 4% of the depression level, metacognitions explained 9% when illness representations, age, gender, history of psychiatric diagnosis, monthly income, and fear of COVID-19 were controlled. The final model of depression, which explained 49% of the total variance, was predicted by average monthly income, illness representations (B-IPQ total score), negative beliefs, and lack of cognitive confidence subscales of MCQ-30. These findings confirm the hypotheses of the study. The MaSC was considered for the first time in T2D. N-MUR made a small but significant contribution to both depression and anxiety levels of patients after illness representations, sociodemographic variables, and fear of COVID-19 were controlled. Moreover, P-MASH was not included in the regression analyses as it was unrelated to the study variables. Also, positive MaSC had a mixed relationship with the impact of FM symptoms (Kollman et al. 2016). These findings require further research to unravel MaSC, particularly positive metacognitions about symptom control.

Although the variance values are low, findings about metacognition are consistent with research showing that metacognitive beliefs predict 33% of anxiety and 27% of depression in people with T2D, after controlling for illness perception, age, and gender (Purewal & Fisher 2017). Consistent with our study, the metacognitive predictor of anxiety in T2D patients was ‘negative beliefs’, and the metacognitive predictors of depression were ‘negative beliefs’ and “lack of cognitive confidence”. Another study showed that the relationship between metacognitive beliefs and rumination was independent of illness representations in T1D and T2D patients (Cherry et al. 2023). Overall, this study was consistent with studies in various physical illnesses, including cancer (Cook et al. 2015a, Cook et al. 2015b), multiple sclerosis (Heffer-Rahn & Fisher 2018), chronic fatigue syndrome (Maher-Edwards et al. 2011), and epilepsy (Fisher & Noble 2017).

The “negative beliefs about uncontrollability of thoughts and danger” have been identified as transdiagnostic metacognition, which was the predictor of depression and anxiety in both T1D and T2D patients (Purewal & Fisher 2017). Similar findings were obtained for this construct in the present study. Moreover, this finding is also supported by a systematic review (Capobianco et al. 2020) that brings together research testing metacognitive beliefs and their relationship with anxiety and depression in physical illnesses. Although studies have highlighted the importance of negative metacognitive beliefs in physical illness, positive beliefs about worry also predicted anxiety in heart and prostate cancer patients after controlling for demographic variables (Anderson et al. 2019). Moreover, the lack of cognitive confidence was a reliable predictor of depression, rather than anxiety, in the current study. This finding supports the studies of Purewal and Fisher (2017).

The present study aims to understand the explanatory power of metacognitions as an independent variable, rather than comparing two theoretical perspectives. While the Self-Regulatory Model (Leventhal et al. 1992, Leventhal et al. 2005) acknowledges illness representations as a significant factor in predicting mood, the S-REF model posits that illness representations may precipitate anxiety or may emerge as a consequence of the anxiety process itself. The results of the present study are consistent with the S-REF model, suggesting that individual responses to illness representations, rather than the representations themselves, are associated with increased emotional distress in patients with T2D (Wells, 2009, Fisher et al. 2018). Moreover, findings on N-MUR can be considered an attentional strategy that seems to support the role of CAS in emotional distress.

Implications for Clinical Practice

Compared to people with diabetes alone, people with diabetes who experience emotional distress are at higher risk of dysfunctions, diabetes complications, and increased healthcare costs. This can make diabetes self-management more difficult and reduce quality of life (Robinson et al. 2023). In this context, it is recommended that, wherever possible, psychosocial treatments should be included in primary healthcare services to improve the management of T2D (Robinson et al. 2023). When considering individual-based psychosocial treatments, systematic reviews and meta-analyses support the effectiveness of motivational interviewing (Bilgin et al. 2022), Cognitive Behavioral Therapy (Li et al. 2017), and Acceptance and Commitment Therapy (Ngan et al. 2021, Sakamato et al. 2022). Focusing specifically on thoughts and representations about the illness appears to be a common feature of these approaches. However, the results of this study suggest that targeting metacognitions, rather than directly targeting illness representations, may be more beneficial for mood regulation in people with T2D. Accordingly, addressing metacognitions as a primary target of intervention has the potential to reduce vulnerability to emotional distress or trauma in chronic illnesses (Cook et al. 2015a). Intervention studies also appear to support this potential effect of metacognitions. In this regard, metacognitive therapy (MCT) is effective in reducing depression (Bayegan et al. 2020, Mirhosseini et al. 2020), fasting blood glucose (Sheikh et al. 2020), body mass index, and emotional eating behaviors (Sheikh, et al. 2021) of T2D patients.

These findings suggest that care-providers, and especially psychiatrists and psychologists, may benefit from MCT to improve the mental health of T2D patients. In addition, training in MCT and getting supervision can help psychologists working with T2D increase their effectiveness. This can also make a positive contribution to health costs. As there is no standardization of the content of MCT in these different studies, this study can contribute to standardizing the content. In light of the findings of this study, focusing specifically on negative metacognitions in therapeutic interventions may be helpful.

Limitations and Future Directions

This study has several limitations. First, the convenience sampling method was employed. This method may have included individuals who were more accepting of their illness in the sample. Second, the study was conducted using a cross-sectional design. As it does not provide data on causality, it is also possible to consider metacognitions as an outcome variable. However, longitudinal studies are required to determine causality and the direction of the relationship. Third, because self-report measurements were used in the study, response bias (social desirability, recall ability) may have increased error variance. Fourth, the study measured only psychological variables related to T2D. However, considering T2D from a biopsychosocial perspective, it is also important to evaluate biological outcomes. Therefore, future research could include measurements such as participants’ body mass index and HbA1c levels to assess glycemic control. Fifth, a significant proportion of patients with T2D (nearly 30%) had been reached online, so a professional could not confirm their diagnosis. Sixth, it is impossible to generalize the study’s data to different types of diabetes, such as gestational diabetes and T1D. In future studies, descriptive and comparative studies can be conducted for different types of diabetes. Seventh, the study was conducted during the COVID-19 pandemic. Although the fear of COVID-19 was controlled in the study, the potential impact of the pandemic on mood should not be ignored. Eighth, illness representations were not assessed as disease-specific since they have not yet been translated into our language. The ‘Diabetes Illness Representations Questionnaire’ (Skinner et al., 2003) may be adapted and used in future studies. Lastly, the emotional distress of T2D patients was limited to depression and anxiety. ‘Diabetes-related distress’ (Martinez et al. 2018), defined as difficulties living with or adapting to diabetes, can be considered in future studies.

CONCLUSION

The S-REF model provides an alternative framework for understanding the underlying mechanisms of emotional distress, including depression and anxiety, in patients with T2D. Beyond illness representations, not only metacognitions but also metacognitions about symptom control contribute to depression and anxiety. Also, experimental and longitudinal studies are needed to examine the causal role of metacognitions in both depression and anxiety. Furthermore, this study, which was conducted for the first time in Turkey, is descriptive and may provide valuable insights into the necessity of examining moderator and mediator variables between illness representations, metacognitions, depression, and anxiety in future studies.